一种利用TM图像自动提取洪积扇的方法

A Method of Alluvial Fan Automatic Extraction from TM Image

在分析洪积扇与其他地物光谱特征差异的基础上,针对TM图像的红光波段与近红外波段的比值能增大洪积扇与其他地物间光谱差异,以及地形阴影在蓝、绿光波段亮度值降低速率差异较大的特征,基于比值运算和差值运算,构建了洪积扇自动提取模型。利用该模型,首先结合阈值算法将洪积扇从其他地物中提取出来,然后采用数学形态学膨胀和腐蚀算法进行空洞填充。在华南丘陵地区的实验表明,该方法除能以较高精度自动提取洪积扇外,还能比较有效地去除植被和阴影等干扰信息。

In this paper, an automatic approach for alluvial fan extraction based on TM image is put forward. Firstly, the spectral feature differences between alluvial fan and other surface features were analyzed;then, in view of the facts that the band ratio between red band and near - infrared band can increase the differences between alluvial fan and other surface features and the decreasing rates of bright values of shadow in blue and green bands are significantly different, the authors built the model of alluvial fan extraction based on ratio and difference operations. Using this model in combination with the algorithm of automatic threshold extraction, the authors separated the alluvial fan from other surface features, and then applied the dilation and erosion filtering algorithm of mathematical morphology. An analysis and comparison of experimental results show that the proposed approach can extract the alluvial fan from hilly areas in South China with high precision. Moreover, the approach can effectively remove the interferential information such as shadow and vegetation.